Write-through vs write-behind caching: understanding the consistency trade-offs

we're designing a new inventory system where data accuracy is absolutely critical. we're trying to figure out the best caching strategy between write-through, write-behind, and write-around. write-through seems safest for consistency because data is written to the cache and then immediately to the database, ensuring the cache is always fresh. however, it adds latency to every write operation. write-behind offers lower write latency, but there's a risk of data loss if the cache node crashes before flushing to the database. write-around is appealing for bulk updates but can lead to cache miss storms immediately after. for an inventory system where every stock update matters, what's the most appropriate strategy? what are the hidden consistency and performance trade-offs you've experienced with each in production?